Welcome to Environmental and Ecological Statistics

  • Environmental & Ecological Statistics is an incredibly broad term covering any form of statistics applied to environmental issues.

  • Key themes include climate change, environmental regulation (eg water and air quality), biodiversity monitoring and ecosystem assessment.

  • This course focuses on this theme rather than a particular type of statistical methodology.

  • We will look at a variety of statistical methods, some of which you will know, and some which will be new.

The Course Structure

Module Week Topic Activities



Environmental Monitoring & Data processing


1 Introduction to Environmental and Ecological Statistics
2 Monitoring and Data mining
3 Sampling and Monitoring Networks Lab 1 + Tutorial 1



Measuring Environmental Change


4 Assessing Change Over Time Tutorial 2
5 Temporal Correlation and Changepoints
6 Modelling Environmental Extremes Lab 2 + Tutorial 3



Environmental and Ecological Spatial Statistics


7 Modelling Areal Data Tutorial 4
8 Modelling Geostatistical Data
9 Methods for Point referrenced Data Lab 3
Special Lecture 10 TBC

Timetable

Lectures: 2 hours per week

Tutorials: 1 hour fortnightly

Practical: 2 hours three times throughout the semester

Assessments

Assessment in this course includes continuous assessment and a final exam. The exam will take place in April/May.

  • Written Exam (75%) - Degree exam in the exam diet

  • Group Report (25%) - Group project

  • Written Exam (65%) - Degree exam in the exam diet

  • Group Report (25%) - Group project

  • Set Exercise (10%) - Critical review of published research

Students must submit at least 75% by weight of the components (including examinations) of the course’s summative assessment.

Questions?

Environmental and Ecological Statistics

  • The environment is (sometimes literally) a burning issue in the 21st century.

Media Coverage

This brings increased focus and interest in statistics as a subject, and how we are working to handle topics like climate change.

BBC article graphs task

  • Choose one of the graphs in the BBC article.

  • Think about what the good and bad aspects (if any) are.

  • Discuss with your neighbour(s) and find out what they thought about their chosen graph.

  • Add some of your thoughts to Mentimeter,

    • E.g. What graph you chose.

    • What is the graph’s purpose?

    • Does it do a good job at serving this purpose?

    • What did you like/dislike about the graph?

Where’s the statistics?

  • Measuring, sampling or monitoring environmental and ecological data, including variation and uncertainty.

  • Ecosystem assessment, detecting and modelling trends, including trends in time and space.

  • Modelling and understanding extreme data.

  • Environmental regulation and policy, and risk assessment.

What are we looking for?

We want to understand changes in the environment and species respond to these changes, in either time, space or both.

  • Are things getting better or worse? Where, when and by how much?
  • What is going to happen next?
  • Where do authorities need to take action, and how can we check if existing actions are working?
  • Also consider complex relationships between environmental variables and species habitats.

Communicating and presenting data

What kinds of data were generated, and how can they be analysed?

Examples

  • Decision making: Is it safe to eat fish from a particular river?
  • Prediction: What is the distribution of the Barn Swallow? Can we predict its distribution in 2030?
  • Regulation: Have emission control agreements reduced air pollutants?
  • Understanding: How did sea levels change over the past 100 years?

Examples: Air pollution

Only one person in ten lives in a city that complies with the World Health Organisation Air quality guidelines.

Examples: Air pollution

Only one person in ten lives in a city that complies with the World Health Organisation Air quality guidelines.

  • The World Health Organisation estimates that 1 in 9 deaths worldwide are due to pollution.
  • The total annual cost of air pollution to the UK economy could be as much as £54 billion.
  • Fine particular matter was associated with an estimated 2,000 premature deaths and 22,500 lost life years in Scotland in 2010.
  • The Cleaner Air for Scotland strategy seeks to reduce air pollution across Scotland.
  • It aims to achieve the “ambitious vision for Scotland to have the best air quality in Europe”

How can we estimate air pollution across Scotland?

Measuring Pollution

  • 99 air quality monitoring stations have been set up across Scotland to capture PM\(_{2.5}\), PM\(_{10}\), NO\(_{2}\), NO\(_{x}\), SO\(_{2}\) and O\(_{3}\).

  • Live data available at

Monitoring Station Map

Estimated PM 2.5 pollution across Scotland

Wildlife monitoring

Monitor wildlife populations to halt biodiversity loss according the to National Biodiversity Strategy and Action Plan (NBSAP).

Wildlife monitoring

Monitor wildlife populations to halt biodiversity loss according the to National Biodiversity Strategy and Action Plan (NBSAP).

  • To halt biodiversity loss we first need to accurately map where species exist and why

Water Quality

Goal 6 of the UN’s Sustainable Development Goals is “Ensure availability and sustainable management of water and sanitation for all”. https://sdgs.un.org/goals/goal6 (UN SDG 6 (clickable))

Water Quality

  • To inform government decision-making, we need to have appropriate data on water quality.
  • We need to sample water quality across lakes and rivers. How do we do this with the resources available?
  • Sampling strategies are required — see the Sampling’ section of the course

Water Quality

  • Once we have the samples, how can we use the data to understand water quality patterns?

  • Statistical modelling approaches are required, often spatial and temporal.

  • We need to understand and report our uncertainties.

Quantification

  • Understanding and measuring quantities is a fundamental part of all science, not just statistics.
  • As scientists, we use data to understand the process which we are investigating.
  • These data have two main sources of uncertainty or error:
    • Inherent variability of the process itself (the thing we are measuring is variable).
    • Imprecise knowledge of the process (our measurements may not be accurate).

Asking questions

  • A big part of our role as statisticians is to ask questions of both our data and our models.
  • How were our data collected? Are they representative of the population? How much uncertainty do we have?
  • Are our models valid? Are the assumptions reasonable? Does the model make sensible predictions? How much uncertainty do we have in our results?
  • These skills are particularly crucial in applied areas such as environmental and ecological statistics.

Example: Arctic Ice

  • Submarines have been used to measure Arctic sea ice.

  • The ice has shrunk both in terms of thickness and extent.

  • We may soon see ice-free summers, which will have a devastating impact on sea life.

  • Arctic sea ice article (clickable)

  • How might we quantify the trend in Arctic ice cover?

Example: Arctic Ice

  • Sometimes, quantification is required.
  • February 2016 sea ice extent was the lowest in the satellite record at 14.22 million square kilometres (5.48 million square miles).
  • The linear rate of decline for February is now 3.0 percent per decade.

Policy, legislation and monitoring

Policy

  • A great deal of environmental statistical research is funded by governments and regulatory bodies.
  • They need to know where the biggest challenges lie so that they can allocate their resources appropriately.
  • Evidence-based policy relies on measuring changes and also evaluating the impacts of existing policies.
  • However, the environment will be one of many competing policy areas, and every government will prioritise it differently.

Policy

  • Policies made in context of multinational agreements.
  • Held annually, COP is the United Nations Climate Change Conference.
  • Aim to limit global temperature rise and achieve net zero emissions.

Policy

  • Environmental policy tends to use very specific language - objectives, targets, guide values, standards, reductions relative to a baseline

  • Policy often prescribes monitoring quantities of interest over space and/or time.

  • Quantities of interest will include water, air and noise pollution, waste management, radioactive substances, biodiversity and animal and plant species

Legislation

Routine Monitoring

  • Monitoring requirements may be prescribed in policy.
  • Much of our data come from routine monitoring of our environmental quantities of interest.
  • Government agencies often make these data available to researchers and/or the public.
  • These data are used to assess compliance with legislation as well as to identify environmental trends and their potential impacts on society.
  • We have already seen monitoring data in the air pollution and water quality examples.

Evidence from routine monitoring

  • Assessment of long-term changes in natural conditions.

  • Assessment of long-term changes resulting from anthropogenic activities.

  • Ascertaining the magnitude and impacts of accidental pollution.

  • Assessing compliance with the standards and objectives of protected areas.

  • Quantifying reference conditions.

Policy, legislation and monitoring

flowchart TB
  A(Policy) --> B(Legislation)
  B --> C(Monitoring)
  C --> D(Analysis)
  D --> A

  • Statistical analyses and reporting is important for ensuring compliance with legislation.
  • Understanding trends, spatial and temporal patterns, and impacts is important for informing policy: evidence-based policy making.
    • E.g. Evaluating effectiveness of current policy, e.g. measurement of change before/after implementation.

Policy, legislation and monitoring

flowchart TB
  A(Policy) --> B(Legislation)
  B --> C(Monitoring)
  C --> D(Analysis)
  D --> A
  

  • Good policy needs a foundation in good science’’ (Wallstr”om, 2005
  • Agencies can be data rich and information poor
  • Other (often short-term) factors play a role in policy making (e.g. news coverage)
  • So, timely findings of practical relevance are important.

Informing policy and legislation

  • What defines “good’’ status?
  • How could we assess this using statistics?

The framework for environmental policy

  • Environmental policy is set at various levels:
    • Internationally (Kyoto Convention, European Union)
    • Nationally, by politicians
  • Environmental policy is implemented and monitored by regulators. In the UK this includes:
  • National Environment Agencies,
  • Government departments.
  • Other parties/stakeholders:
    • Industrialists (extra cost), economists, special interest groups and organisations (Greenpeace, WWF etc.), the public.

Summary points

  • Environmental and Ecological statistics is a broad term covering many different techniques.

  • It can involve:

    • Decision-making
    • Prediction
    • Regulation
    • Understanding
  • We need to communicate and present data and statistics to e.g. the public, government and subject-matter experts.

Summary points

  • To understand quantities, we need to understand the process.
  • We need to quantify the uncertainty and error, which have multiple sources:
    • Inherent variability of the process.
    • Imprecise knowledge of the process.
  • We must collaborate with subject-matter experts, to ensure that we understand e.g. how the data were collected, are our model’s predictions sensible?

Summary points

  • We need to understand the policy and legislative context:
    • International agreements \(\rightarrow\) National-level policies \(\rightarrow\) Legislation.
    • Legislation is the legal framework used to implement policy.
    • Regulatory bodies monitor compliance with legislation through collecting routine monitoring data.
  • Our work can inform policy decisions through ensuring that representative data are collected, and that results are appropriately interpreted and understood.

References

  • Piegorsch, W. W., & Bailer, A. J. (2005). Analyzing environmental data. Wiley. (Available from the University Library as an e-book here ).

  • Barnett, V. (2004). Environmental statistics: Methods and applications. Wiley. (Available from the University Library as an e-book here ).

  • Manly, B. F. J. (2001). Statistics for environmental science and management. Chapman & Hall/CRC. (No e-book available, but a physical copy is available from the University Library ).